[b-hebrew] Masoretic invention, was: masorete pointing v's LLX transliterations

Yitzhak Sapir yitzhaksapir at gmail.com
Tue Jul 26 14:47:02 EDT 2005

Dr. Joel M. Hoffman wrote:
> >The Masoretes would not dream of changing anything or invent 
> >anything new.
> Not on purpose.  But there is incontrovertible evidene that the
> Masoretes did not correctly capture the sounds of ancient 
> Hebrew.
> The clearest example comes from the Masoretic invention of 
> dropping a Dagesh after a word that ends in a vowel sound and 
> that is pointed with a conjunctive trope.  The reason this system 
> cannot reflect any real dialect of Hebrew is that the conjunctive 
> tropes do not play a uniform role in grammar: Sometimes the 
> exact same grammatical relationship will be represented by 
> different tropes, and sometimes one of those tropes is 
> conjunctive while another is disjunctive.  A handful of these 
> situations involves a dagesh in the following word.
> For example, Judges 1:1 and 1:8 contain the same syntactic structure,
> but exhibit different patterns of trope and Dagesh:
>     Judges 1:1  ...and the children of Israel inquired...
>     Judges 1:8  ...and the children of Judah fought...
> In Judges 1:1, "fought" gets a disjunctive trope, while in 1:8, it
> gets a conjunctive trope.  Because of this difference, BNEY 
> (YISRAEL or YEHUDAH), which follows the verb in Hebrew, gets
> a Dagesh only Judges 1:1, but not in 1:8, even though the 
> syntactic structure is exactly identical.  This cannot represent a 
> spoken dialect of Hebrew.

Hello Dr. Hoffman,

According to Wikipedia, the ta marbuta in Arabic is pronounced
as -t- only at "pausa", and not "sandhi".  This accords with 
something I learned in the past, but apparently, in actual
speech it behaves more like Hebrew which makes sense since
this is likely a phonological rule in Western Semitic.  In any case,
the fact of the matter is that certain phonological rules are active
only at pausal locations, ie, when no word immediately follows it.
A "conjuctive trope" is one that exhibits the smallest amount of
stop between words.  There are four different degrees of
"disjunctive tropes" but all of them exhibit longer pauses.  This
is used in grammar in an indirect way.  In both the above 
sentences, the section "And the children of Israel/Judah
inquired/fought" is a single unit in the verse.  Within this single
unit, we have the noun which forms one unit, and the verb.
The noun itself is composed of two parts - "children of" and 
"Israel/Judah".  In Jud 1:8, though, "children of" is attached
to the following word using a maqqaf, and therefore forms a 
single unit.  So the pause between the verb and the noun need
not be great.  A conjuctive trope is enough.  In Jud 1:1, the two 
are separate words, so that to join the two parts of the noun, 
there is a conjuctive trope under "children of", a munnax.  In order 
to show that "children of Israel" is a single unit, a higher pausal
degree than the munnax must come between this unit and the
verb.  This brings it into the next higher degree of disjunction, and
consequently, there is a disjunctive trope.  As a side effect, since
there is now a longer pause in speech between the two words
("fought" and "children"), the word "children of" acts slightly 
different and receives a dagesh qal.  In Jud 1:8, this didn't happen
because a dagesh qal doesn't come after an open syllable and the 
pausal strength of the conjuctive trope is not that great to undo
the open syllable.

And now comes the question, why is there a maqqaf in one, and
not in the other?  Well, it's possible that it's just accident that
in some manuscript in one the two were written together as one 
word and in the other there was a small space between the two
words.  In reading the words, now, the Massoretes therefore read
the first as one word, and the second as two words and hence had
to add a slightly higher degree of pause in the second case.

What is clear, however, is that all of this is a phonological rule
and it is odd that you are arguing that a phonological rule has
no basis in speech.  While I think Wikipedia is wrong on the
ta marbuta issue in Arabic, it is clear that there are phonological
rules that take effect only dependent on the variant degree of
pause between the two words (which Wikipedia refers to as
pausa and sandhi).  Rather than claiming that because of such 
phonological rules the Massoretic rendering is artifical, I think
one should rather hold the opposite.  The Massoretic rendering
is based in spoken Hebrew and very accurately catches the
sound of the spoken Hebrew.

Yitzhak Sapir

Wikipedia links:

More information about the b-hebrew mailing list